Dear Dolores
I am not familiar with this user-written command, but it is never
possible to use non-integer frequency weights in any context.
Analytical weights may be non-integer (but must be positive, I
believe).
It appears that the ginidesc routine has an error in its handling of
weights, in that if one specifies analytical weights, it indicates that
they are being treated as frequency weights:
. ginidesc invest [aw=lk], by(company)
may not use noninteger frequency weights
r(401);
.
I believe this is being caused by lines such as
ineqdeco `inc' [w = `wi'] if `bb'==`i'
qui ineqdeco `inc' [w = `wi']
in which the weight type is not being specified. It appears that
ineqdeco's "natural" weight type is fw, so that if not specified, it
takes the weights as fws. This is not ineqdeco's fault; it is not being
passed the information as to weight type.
You might modify a copy of ginidesc, and/or contact the authors of
ginidesc and encourage that they release a corrected version of the
program.
Best wishes
Kit
On Apr 23, 2004, at 11:33 AM, D Jimenez Rubio wrote:
Dear Prof. Baum,
I am using the stata command ginidesc. However, I get an error when
trying to set weights in the analysis. Specifically, what I write in
Stata is:
ginidesc hui [fw=iwt] , by(incgr) m(a1) gk(a2)
However, I get the following error:
may not use noninteger frequency weights
r(401);
I get the same error message even if I set analytical weights instead
of
frequency ones.
I would extremely appreciate any suggestions concerning this problem.
Thanks a lot in advance. Sincerely,
Dolores Jimenez Rubio
University of York (UK)
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
